| Description |
Current comparison methods for experimental and simulated holographic interferometric images are qualitative in nature. Previous comparisons of holographic interferometric images with computational fluid dynamics (CFD) simulations for validation have been performed qualitatively through visual comparison by a data analyst. By validating the experiments and CFD simulations in a quantifiable manner using a consistency analysis, the validation becomes a repeatable process that gives a consistency measure and a range of inputs over which the experiments and CFD simulations give consistent results. The quantification of uncertainty in four holographic interferometric experiments was performed for use in a data collaboration with CFD simulations for the purpose of validation. The model uncertainty from image-rocessing, the measurement uncertainty from experimental data variation, and the scenario uncertainty from the bias and parameter uncertainty was quantified. The scenario uncertainty was determined through comparison with an analytical solution at the helium inlet (height, x = 0), including the uncertainty in the experimental parameters from historical weather data. The model uncertainty was calculated through a Box-Behnkin sensitivity analysis on three imageprocessing code parameters. Measurement uncertainty was determined through a statistical analysis to determine the time-average and standard deviation in the interference fringe positions. An experimental design matrix of CFD simulations was iv performed by Weston Eldredge using a Box-Behnkin design with helium velocity, temperature, and air co-flow velocity as parameters in conjunction to provide simulated measurements for the data collaboration Data set. Over 3,200 holographic interferometric images were processed through the course of this study. When each permutation of these images is taken into account through all the image-processing steps, the total number of images processed is over 13,000. Probability distribution functions were plotted for each interference fringe order at each measurement height, making a total of 22 PDFs. Model, scenario, and measurement uncertainty was quantified in the experiments. The CFD simulations were performed. The final uncertainty attributed to the experiments resulted in a maximum uncertainty of -7.96 fringes. The largest contributor to uncertainty was the scenario uncertainty with measurement uncertainty as the second largest. The model uncertainty was very small and as such had the smallest contribution to the overall experimental uncertainty. In the future, the results of this study will be used in conjunction with the CFD simulations discussed and their attributed error in a data collaboration to determine Data set consistency for validation. |
